### Table 7.9: Walkthrough compared on a PC 100, Mesa library.

### Table 7.10: Walkthrough compared on a SGI Onyx, Mesa library.

### Table 3. Total rendering time in sec for five 3D data sets using different number of processors.

1998

"... In PAGE 5: ...aragon node. Table 2 shows these measurements. To evaluate the performance of our parallel library, we wrote a user program that reads in 3D datasets (triangles and triangle strips) from disk, performs data distribution, and makes calls to our parallel library to render and display the image. Table3 and Figure 4 show the performance of parallel Mesa in terms of total rendering time and the triangle ren- dering rate to generate the final images with varying num- ber of processors. With increasing number of processors, the total rendering time is decreasing and the rendering rate is increasing.... ..."

Cited by 5

### Table 3.1: Parallel search libraries.

### Table 1. MPI and SHMEM library calls used in the parallelization.

"... In PAGE 4: ... The WorkPool algorithm proved to be a suitable solution for coverage calculation. Implementation Table1 presents the MPI [1,6] and SHMEM [2, 3] library calls which are used for parallel approach of the coverage calculation. MPI_Init() and MPI_Finalize() calls are necessary for the initialization and termination of the parallel environment.... ..."

### Table 2. Selected Nested Sequence Operations in Data Parallel Library

1996

"... In PAGE 9: ...class of operations which have a one-to-one correspondence between consumption of input pieces and generation of output pieces, such as elementwise operations, are called participants. Table2 provides a list and categorization of several representative data- parallel primitive operations. Name Action Piecewise Behavior arith-ops basic arithmetic and logical operations participant substitute replaces every sequence element with a supplied value participant distribute replicate values to form a sequence generator range1 enumerate integers between 1 and a supplied value generator length the number of elements in a sequence accumulator x reduce family of reduction operations (+, , and, or, max, min) accumulator... ..."

Cited by 9

### Table 2. Selected Nested Sequence Operations in Data Parallel Library

1996

"... In PAGE 9: ...class of operations which have a one-to-one correspondence between consumption of input pieces and generation of output pieces, such as elementwise operations, are called participants. Table2 provides a list and categorization of several representative data- parallel primitive operations. Name Action Piecewise Behavior arith-ops basic arithmetic and logical operations participant substitute replaces every sequence element with a supplied value participant distribute replicate values to form a sequence generator range1 enumerate integers between 1 and a supplied value generator length the number of elements in a sequence accumulator x reduce family of reduction operations (+, , and, or, max, min) accumulator... ..."

Cited by 9

### TABLE I CHARACTERISTICS OF DIFFERENT LIBRARIES RELATED TO PARALLEL VSIPL++.

2004